“…As the field of machine learning has evolved over the last decade, these algorithms have been used to train models for predicting various outcomes based on EDA data or multi-sensor data including EDA. Examples from works published within the last few years include classification of epileptic seizures (Zsom A and et al, 2019), differentiation of sensory responses for children with autism spectrum disorder (Raya MA et al, 2020), identification of cognitive tasks (Posada-Quintero HF and Bolkhovsky JB, 2019), pain assessment (Susam BT et al, 2018;Posada-Quintero HF et al, 2021;Aqajari SAH et al, 2021), emotion recognition (Al Machot F et al, 2019;Sharma V et al, 2019;Ganapathy N et al, 2020), assessment of emotional engagement (Di Lascio E et al, 2018), Stress detection (Amalan S and et al, 2018;Zontone P et al, 2019;Pakarinen T et al, 2019;Anusha AS et al, 2020;Sánchez-Reolid R et al, 2020;Greco A and et al, 2021), cognitive load measurement (Romine WL et al, 2020), detection of major depressive disorder (Kim AY et al, 2018), and arousal detection from music ( Bartolomé-Tomás A et al, 2020). Among these studies, high performances in classification accuracies have been reported ranging from 64% to 95% depending on the dataset, classification problem and difficulty A c c e p t e d M a n u s c r i p t (binary classification is for instance less challenging than multiple levels of stress).…”